Heat Controller stop regulating

[meute 2]

Hi,

I have just put in service 8 "Heat Controller" with external temp sensor and the 2 or 3 first days the regulation was working, not very well, with big overshot and abnormal temperature variations ... but the valve motor was reacting to setpoint change and room temp change.

 

Today, since this morning, the 8 valve are not regulating anymore, if I force it to close > they close, if I force it to open > they open, but if I give a setpoint in "HEAT" mode the valves stay closed (for hours) and the setpoint is 3 or 4 °c higher than the measured room temperature.

 

The window opened sensor mode is completely disabled.

 

Calibrating again changes nothing, going to standby mode and come back to working mode changes also nothing.

 

I don't know how to make it working again, for sure a factory reset should help but in 2 or 3 days I will get the same problem, and if I made a factory reset I must reconfigure all that part in my Home controller system. (I use Jeedom).

 

Also it's not admissible that the only solution to upgrade the firmware is to use a fibaro HC ... If you sell modules with firmware bugs you must provide solution to update it to all customers and not only for the customers who use your HC. With special tool by USB or all other possible solutions. 

Edited January 11, 2018 by meute

[Gianluca S. 1]

6 hours ago, petergebruers said:

 

He probably can't, because he is not using a Home Center (it is a Jeedom)... But you can try my approach posted here:

 

 

 

Thanks. I have red your Lua code.

Question: If the radiator reach the desired temperature, close the valve?

 

Seems that the firmware will be delivered in March (source: Fibaro support)...

...so we have to manage is some way this time period.........

 

regards

Gianluca

Edited January 17, 2018 by Gianluca S.

[petergebruers 1,268]

3 hours ago, rhutin91 said:

My valves are also slow to react.

I am intrigued. If you find the time, can you elaborate to satisfy my curiosity? Can you define "slow". Slow to respond to environmental conditions? Slow to heat the room? Do you have some data to share?

 

I think, with the issue of the current firmware, it is hard to tell how good it will perform after the firmware update. That is why I haven't published a lot of data myself. So I perfectly understand why you would not want to discuss this...

 

2 hours ago, meute said:

I'm pretty sure that Fibaro will correct the problem and will finally do an auto-tuned PID loop who is working ... others do it with success.

 

I understand what you are saying and I fully agree. If someone wants to implement a PID, based on some temperature measurement device, please use forum "search". Someone published functional Lua code on this forum. I do not use it, but I think it can be used to control an FGT.

 

Edit: as @meute pointed out, running a PID script on your -Zwave controller to control your FGT-001 won't work, it won't work any better than simple on-off control. I fully agree with him.

Edited January 17, 2018 by petergebruers

[rhutin91 0]

33 minutes ago, petergebruers said:

I am intrigued. If you find the time, can you elaborate to satisfy my curiosity? Can you define "slow". Slow to respond to environmental conditions? Slow to heat the room? Do you have some data to share?

 

I think, with the issue of the current firmware, it is hard to tell how good it will perform after the firmware update. That is why I haven't published a lot of data myself. So I perfectly understand why you would not want to discuss this...

 

 

I understand what you are saying and I fully agree. If someone wants to implement a PID, based on some temperature measurement device, please use forum "search". Someone published functional Lua code on this forum. I do not use it, but I think it can be used to control an FGT.

 

I have currently converted 8 of my 20 valves to Fibaro. They are connected to an eedomus+ box. They are all slow (typically 2 to 3 hours) to heat the room when the set point is changed. My system is a central water heater with the water temperature based on outside temperature and set point of the dining room. In the short term , I am anticipating the set point changes to get the correct temperature on time. 

With my old valves the heating took about 1/2 hour. The old system was based on the german FS20 protocol with Conrad valves. But it was very difficult to get a reliable remote control.

 

I fully agree that a firmware update may fix the problem. I am waiting for it before developping my own FID script in php for eedomus. (I have a fibaro home lite just for the fibaro module update. 

[meute 2]

10 minutes ago, petergebruers said:

If someone wants to implement a PID, based on some temperature measurement device, please use forum "search". Someone published functional Lua code on this forum. I do not use it, but I think it can be used to control an FGT.

A good PID on a slow response process requires a proportional actuator but not all-or-nothing actuator and so far we can only force the opening or closing of the valve completely. Without this, the valve will probably perform the entire stroke too frequently.

 

The proportional action PID will preserve the battery life by making mostly small adjustments and large adjustments only when necessary, mostly after a setpoint change.

 

With an all-or-nothing action and a high inertia process like home heating, the best way is to work with a simple and easy to implement hysteresis control that will reduce the frequency of valve motion and preserve the life of the battery.

And that gives a result good enough to wait for the update.

 

42 minutes ago, Gianluca S. said:

Seems that the firmware will be delivered in March (source: Fibaro support)...

Really, what a pity. so so long time ...

 

I think that the problem for fibaro to release faster publicly something is all the customers who can't updates the module themselves and need thus to send it back to the reseler to use the warranty circuit ...

If all users was using it only with a Fibaro Home Center we will get beta firmware faster. With this additional constraint they must be sure at 200% that the update corrects all the problems in one shot and it's taking a lot of time (Thing who should have be done normally long time before).

 

Also announce a battery life of "one heating season" is easy with something who never regulate anything and stay opened or closed most of the time ... but with something who make a correct regulation and thus move most frequently the promise is surely harder to honorate ...

[petergebruers 1,268]

2 hours ago, rhutin91 said:

They are all slow (typically 2 to 3 hours) to heat the room when the set point is changed.

 

Thank you for explaining this to me... I do not get such long times. Heating in my TV room set to 21 degrees, start temp = 18 degrees. Target reached in about 1/2 hour.

 

2 hours ago, meute said:

A good PID on a slow response process requires a proportional actuator but not all-or-nothing actuator and so far we can only force the opening or closing of the valve completely. Without this, the valve will probably perform the entire stroke too frequently.

 

Ha, yes, when I wrote about PID control I should have thought a bit longer.  Of course you are right. It is probably not going to work better than a simple on-off control, can't beat the laws of physics. I was thinking you could kind of PWM the valve, say 10 minute period and 1 minute intervals to get 10 % steps just for fun. But that would drain the battery and make a lot of noise. I am thinking out loud. I am not even going to try that. I'm not going to use my HC for any type of temperature control if I can avoid that... 

 

Edited January 17, 2018 by petergebruers